AIGR Gene Hypothesis Deep Research — Final Report

Target: Schizosaccharomyces pombe mmf2 (Q9UR06 / SPAC1039.10)

Seed hypothesis: S. pombe mmf2 is a mitochondrial Rid-family enzyme whose specific molecular function is 2-iminobutanoate/2-iminopropanoate (including 2-aminoacrylate) deaminase activity (GO:0120241), rather than only generic deaminase activity (GO:0019239).

Focus type: core_function

Summary

The seed hypothesis is supported on homology and active-site grounds. S. pombe mmf2 (Q9UR06, 126 aa) carries every diagnostic feature of a catalytically active RidA-subfamily imine deaminase: it is assigned InterPro IPR006056 (RidA) and IPR006175 (YjgF/YER057c/UK114), Pfam PF01042 (Ribonuc_L-PSP), and PANTHER PTHR11803 — a family whose canonical name is literally "2-iminobutanoate/2-iminopropanoate deaminase RidA." UniProt's "RutC family" wording is legacy nomenclature for this same YjgF/RidA fold, not evidence of a distinct RutC subfamily activity. A global BLOSUM62 alignment computed in this investigation places mmf2 Arg101 exactly on the E. coli RidA catalytic Arg105 — the invariant residue essential for imine hydrolysis that is present only in the active RidA/Rid1–Rid3 subfamilies (absent in the inactive Rid4–Rid7 subfamilies). mmf2 is therefore in the active branch.

The closest characterized ortholog, S. cerevisiae Mmf1p (same PANTHER subfamily PTHR11803:SF58 "HMF1-RELATED"), is a directly demonstrated mitochondrial 2-aminoacrylate/enamine–imine deaminase. Independently, the human ortholog HRSP12/UK114 (P52758) is already annotated with the two children of GO:0120241 (GO:0120242 via TAS:Reactome; GO:0120243 via IEA:RHEA), establishing that the GO consortium already treats the specific deaminase terms — not the generic GO:0019239 — as the appropriate molecular function for this family. Taken together, GO:0120241 is a better-justified core molecular function for mmf2 than the current generic GO:0019239.

The single most important caveat is that no direct enzymatic assay exists for Q9UR06/SPAC1039.10 itself — the assignment rests on family membership plus active-site conservation plus ortholog data. The correct evidence code for adding GO:0120241 is therefore ISS/ISO (with/from Mmf1p P40185 and/or human HRSP12 P52758), not an experimental code. Additionally, "mitochondrial DNA maintenance" and related growth/heme phenotypes observed in yeast orthologs are downstream biological-process consequences of the deaminase activity, not the core molecular function, and should be flagged as non-core.

Key Findings

Finding 1 — mmf2 is a RidA-subfamily enzyme retaining the catalytic arginine; its specific MF is GO:0120241

Q9UR06 is a 126-amino-acid protein annotated by UniProt with InterPro IPR006056 (RidA) and IPR006175 (YjgF/YER057c/UK114), Pfam PF01042 (Ribonuc_L-PSP), and PANTHER PTHR11803, named "2-iminobutanoate/2-iminopropanoate deaminase RidA" (subfamily PTHR11803:SF58, "PROTEIN HMF1-RELATED"). This domain architecture is the defining molecular signature of the Rid superfamily.

Membership in the catalytically active branch of the family is established not by domain assignment alone but by the presence of a conserved active-site arginine. Niehaus et al. (PMID: 25975565) established the diagnostic criterion: "The Rid1 to Rid3 subfamilies, but not the Rid4 to Rid7 subfamilies, have a conserved arginine residue that, in RidA proteins, is essential for imine-hydrolyzing activity." A BLOSUM62 global alignment computed in this investigation places mmf2 Arg101 aligning exactly to E. coli RidA catalytic Arg105, placing mmf2 firmly in the active RidA branch rather than a degenerate subfamily.

The mechanistic activity underlying GO:0120241 is well defined. Lambrecht et al. (PMID: 22094463) demonstrated that "YjgF proteins have enamine/imine deaminase activity and accelerate the release of ammonia from reactive enamine/imine intermediates of the pyridoxal 5'-phosphate-dependent threonine dehydratase (IlvA)." This is exactly the reaction described by GO:0120241.

The closest characterized ortholog, S. cerevisiae Mmf1p, is a mitochondrial member of this family: "The mitochondrial RidA homolog, Mmf1p, prevents enamine/imine stress and is important for normal growth and maintenance of mitochondrial DNA" (PMID: 34559917). This confirms both the mitochondrial localization and the enamine/imine deaminase activity of the closest ortholog, and frames mtDNA maintenance as a downstream consequence of the deaminase activity rather than the core function.

Pairwise identity of mmf2 to characterized RidA-family members is in the expected homolog range (~30–43%): Hmf1p ≈43%, Mmf1p ≈34–42%, E. coli RidA ≈30–37%, human UK114/HRSP12 ≈34–49%. The current UniProt/PomBase GO MF for mmf2 is only the generic GO:0019239 (deaminase activity, IBA), which is less informative than the family-defining specific term.

Finding 2 — GO:0120241 is the family-defining MF for the RidA/PTHR11803 group; the human ortholog is already annotated with its children, and S. pombe has two RidA paralogs (mmf1, mmf2)

QuickGO defines GO:0120241 as "Catalyzes the hydrolytic deamination of imine intermediates formed by several types of pyridoxal-5'-phosphate-dependent dehydratases, such as EC 4.3.1.19 and EC 4.3.1.17" — i.e., exactly the RidA activity. It has two children: GO:0120242 (2-iminobutanoate → 2-oxobutanoate) and GO:0120243 (2-iminopropanoate → pyruvate).

The human ortholog HRSP12/UK114 (P52758) is already annotated with these specific child terms: GO:0120242 (TAS:Reactome) and GO:0120243 (IEA:RHEA). This is a strong term precedent — the GO consortium already treats the specific deaminase terms, not the generic GO:0019239, as the appropriate molecular function for this family.

A UniProt taxonomy search (taxid 284812, InterPro IPR006175) returns exactly two RidA-type paralogs in S. pombe: mmf1 (O43003, 162 aa) and mmf2 (Q9UR06, 126 aa). Both are PANTHER PTHR11803:SF58 "HMF1-RELATED," both are annotated "RutC family," both localize to mitochondrion + cytosol, and both currently carry only generic GO:0019239. (Note: mug71/Q9USQ7 shares the fold reference but is a diphthine-ammonia ligase — an unrelated activity — and should not be conflated with the RidA paralogs.)

Finding 3 — The catalytic arginine (E. coli RidA Arg105 equivalent) is conserved in mmf2 and every RidA homolog tested

A global Needleman–Wunsch alignment (BLOSUM62, gap = −6) of each candidate protein against E. coli RidA (P0AF93), mapping to reference position 105, gave the following results:

The essential catalytic arginine — "a conserved arginine residue that, in RidA proteins, is essential for imine-hydrolyzing activity" (PMID: 25975565) — is present in mmf2 and in every homolog tested, including both S. pombe paralogs. This confirms catalytic competence at the sequence level: both fission-yeast paralogs are in the catalytically active RidA branch, not a degenerate Rid4–7 subfamily. Computed provenance is preserved in artifacts/mmf2_catalytic_arg_conservation.csv.

Mechanistic Model / Interpretation

The Rid/RidA reaction and mmf2's place in it:

   PLP-dependent dehydratases
   ├─ Threonine dehydratase (IlvA, EC 4.3.1.19)  ── Thr/Ile pathway
   └─ Serine/cysteine dehydratase (EC 4.3.1.17)  ── Ser/Cys pathway
      │  (α,β-elimination)
      ▼
   2-iminobutanoate  /  2-iminopropanoate (= 2-aminoacrylate, 2AA)
      │        ← reactive enamine/imine; METABOLIC STRESSOR
      │           (covalently inactivates PLP enzymes)
      │
      │   RidA-family deaminase  ◄── mmf2 acts HERE (GO:0120241)
      │   (catalytic Arg101 + active-site water)
      ▼
   2-oxobutanoate  /  pyruvate   +   NH3   ← stable, harmless products

Direct molecular function (core, what to annotate): mmf2 is proposed to catalyze the hydrolytic deamination of reactive 2-imino/enamine intermediates — 2-iminobutanoate (from Thr/Ile metabolism via threonine dehydratase, EC 4.3.1.19) and 2-iminopropanoate/2-aminoacrylate (from Ser/Cys metabolism, EC 4.3.1.17) — yielding 2-oxobutanoate/pyruvate plus ammonia. This is GO:0120241 and its children GO:0120242/GO:0120243.

Downstream biological process (not core MF): By deaminating these reactive intermediates before they covalently inactivate PLP-dependent enzymes, RidA-family proteins protect central metabolism. In mitochondria, the consequence of losing this activity (documented for yeast Mmf1p) is enamine/imine stress, impaired growth, disrupted heme biosynthesis (via Hem1p), and loss of mitochondrial DNA. These are pleiotropic downstream phenotypes, not the immediate catalytic function.

Cellular location: Both S. pombe RidA paralogs are annotated mitochondrion + cytosol; the closest characterized ortholog Mmf1p is mitochondrial. The seed hypothesis's "mitochondrial" framing is consistent with this, though the more complete current annotation is dual localization (mitochondrion + cytosol).

The curation logic chain:

Domain (RidA/PTHR11803)  +  conserved catalytic Arg101  +  characterized ortholog Mmf1p
└──────────────────────────┬──────────────────────────┘
                    ▼
   GO:0120241 (2-iminobutanoate/2-iminopropanoate deaminase activity)
   assigned by ISS/ISO  (with/from: Mmf1p P40185 or human HRSP12 P52758)
                    │
                    ▼
   Downstream (non-core):  enamine/imine detoxification, metabolic-stress
           protection, mtDNA maintenance = phenotypes

Evidence Base

How the papers connect: PMID: 25975565 supplies the sequence criterion (catalytic Arg) that lets us place mmf2 in the active branch; PMID: 22094463, PMID: 23386433, PMID: 40596262, and PMID: 38801230 supply the direct biochemistry of the family reaction (GO:0120241); and PMID: 34559917 plus PMID: 29487232 characterize the closest ortholog Mmf1p and demonstrate that mtDNA maintenance is downstream. PMID: 32576850 provides a cautionary note that paralogs can diverge biochemically — relevant to mmf1 vs mmf2.

(Computed provenance: artifacts/mmf2_catalytic_arg_conservation.csv.)

GO Curation Implications (leads — require curator verification)

• ADD MF GO:0120241 (2-iminobutanoate/2-iminopropanoate deaminase activity) with evidence code ISS/ISO (with/from S. cerevisiae Mmf1p P40185 and/or human HRSP12 P52758), or retain IBA if the GO_Central IBA set is refreshed. GO:0120241 is defined as the "hydrolytic deamination of imine intermediates formed by … PLP-dependent dehydratases such as EC 4.3.1.19 and EC 4.3.1.17" — i.e., exactly the RidA activity. This directly answers the seed hypothesis: the specific term is warranted.
• Do NOT use an experimental code (IDA/IMP/EXP) — there is no direct assay on Q9UR06.
• Generic GO:0019239 (deaminase activity) is too broad; it should be superseded by GO:0120241 (or retained only as the redundant parent).
• Optionally add the two children GO:0120242 / GO:0120243, mirroring the human ortholog HRSP12; but the parent GO:0120241 is the safest single-term choice absent mmf2-specific substrate data.
• CC: retain mitochondrion (e.g., GO:0005759 mitochondrial matrix, ISO) and cytosol (GO:0005829) — dual localization matches both paralogs and the predicted N-terminal transit peptide.
• Reclassify "mitochondrial DNA maintenance" free-text/annotation as non-core / downstream BP phenotype, not the molecular function (PMID: 29487232 shows it is a consequence of unmanaged 2AA).
• Avoid "protein binding" — a specific, well-supported catalytic MF is available and far more informative.

Mechanistic Scope

Immediate molecular function (core): hydrolytic deamination of reactive 2-imino acids — 2-iminobutanoate (from Thr/Ile metabolism, EC 4.3.1.19) and 2-iminopropanoate/2-aminoacrylate (from Ser/Cys metabolism, EC 4.3.1.17) — yielding 2-oxobutanoate/pyruvate + NH₄⁺. This is the direct gene-product activity captured by GO:0120241.

Downstream effects (not core):
- Prevention of 2-aminoacrylate damage to PLP enzymes — the protective consequence of the deaminase activity.
- Mitochondrial DNA maintenance (yeast Mmf1p) — a downstream consequence of preventing enamine/imine stress.
- Heme biosynthesis (via Hem1p, PMID: 34559917) — pathway-level downstream effect in yeast.
- Growth defects / respiratory competence / isoleucine biosynthesis — loss-of-function phenotypes.

None of these should be annotated as the core molecular function; they are properly BP/phenotype-level observations.

Conflicts and Alternatives

1. Paralog/ortholog confusion (principal caveat). All direct enzymatic evidence is from Mmf1p, Hmf1p, bacterial RidA, RutC, human HRSP12, and fish/goat/streptococcal RidA — not from Q9UR06. The assignment for mmf2 is inference by homology, appropriately coded ISS/ISO.

2. Paralog substrate divergence. PMID: 32576850 shows that gene-duplication paralogs within the Rid family can diverge biochemically. S. pombe has two RidA paralogs (mmf1, mmf2), both retaining the catalytic Arg. This does not undermine the family assignment but means the precise child term (GO:0120242 vs GO:0120243) for mmf2 specifically is not established, and there is no direct evidence of narrowed/divergent specificity between the two paralogs. GO:0120241 (the generalist parent) therefore applies to both.

3. PANTHER subfamily / "RutC family" wording. PTHR11803:SF58 is named "PROTEIN HMF1-RELATED," and UniProt uses "RutC family" language. This is legacy nomenclature for the YjgF/RidA fold, not evidence of a distinct RutC pyrimidine-utilization activity. Because mmf2 retains the conserved catalytic Arg and its closest characterized ortholog (Mmf1p) is a bona fide 2AA deaminase, the "HMF1-related/RutC-family" wording does not argue against GO:0120241 — it reflects naming history.

4. Legacy RNase / translation-inhibitor annotations. The UK114/p14.5 family historically carried "endoribonuclease" and "translation inhibitor" annotations. These are widely regarded as secondary or artifactual since the RidA deaminase activity was established and should not be propagated to mmf2.

5. Rid4–Rid7 inactive subfamilies. These lack the catalytic Arg and would NOT justify GO:0120241. mmf2 is explicitly not in this group — it retains Arg101 — so this alternative is ruled out by the sequence analysis.

Limitations and Knowledge Gaps

Additional limitation: the inference is homology/active-site based; PubMed access via the provided tool was intermittent for some queries; and localization for mmf2 is by similarity (ISO/ECO:0000250), not measured directly in fission yeast.

Discriminating Tests

1. Heterologous complementation: Express mmf2 in an S. enterica ridA⁻ strain (or S. cerevisiae mmf1Δ) and test rescue of 2AA-mediated stress / growth on minimal medium — the classic decisive functional test (used for SsRidA in PMID: 40596262).

2. In vitro deaminase kinetics: Purify recombinant mmf2 and measure deamination of 2-aminoacrylate (→ pyruvate) and 2-iminobutanoate (→ 2-oxobutanoate) via the coupled L-amino-acid-oxidase or IlvA/IlvE assay. Distinguishes GO:0120242 from GO:0120243 preference.

3. Active-site Arg mutagenesis: Generate mmf2 R101A and confirm loss of activity — directly tying the conserved arginine identified computationally to catalysis (mirrors goat RidA, PMID: 38801230).

4. Localization: GFP-tagged mmf2 in S. pombe to confirm mitochondrial and/or cytosolic localization.

5. Genetic interaction / paralog dissection: mmf1Δ, mmf2Δ, and mmf1Δmmf2Δ under serine/cysteine/isoleucine-pathway stress plus metabolomics for 2AA-adduct accumulation and mtDNA/respiration phenotypes.

Proposed Follow-up Actions / Curation Leads (verify before applying)

• Candidate new MF term: GO:0120241 (ISS/ISO, with-from P40185 and/or P52758). Optional children GO:0120242, GO:0120243.
• Action change: downgrade/replace generic GO:0019239; reclassify "mtDNA maintenance" free text as a downstream phenotype rather than core function.
• Candidate references + snippets to verify:
• PMID: 34559917 — "The mitochondrial RidA homolog, Mmf1p, prevents enamine/imine stress and is important for normal growth and maintenance of mitochondrial DNA."
• PMID: 22094463 — "YjgF proteins have enamine/imine deaminase activity and accelerate the release of ammonia from reactive enamine/imine intermediates of the pyridoxal 5'-phosphate-dependent threonine dehydratase (IlvA)."
• PMID: 25975565 — "The Rid1 to Rid3 subfamilies, but not the Rid4 to Rid7 subfamilies, have a conserved arginine residue that, in RidA proteins, is essential for imine-hydrolyzing activity."
• Suggested curator questions: Is an ISS MF acceptable given no pombe-specific assay? Should both paralogs (mmf1, mmf2) receive the same MF? Retain generic GO:0019239 as redundant parent or remove?
• Suggested experiments: the discriminating tests above, prioritizing heterologous complementation and R101A mutagenesis.

Conclusion

The seed hypothesis is supported: mmf2 has every hallmark of an active RidA-subfamily enzyme — the RidA/PTHR11803 domain assignment plus the conserved catalytic Arg101 (aligning to E. coli RidA Arg105) — and its closest characterized ortholog (yeast Mmf1p) is a demonstrated mitochondrial 2-aminoacrylate/enamine–imine deaminase. GO:0120241 is a better-justified core molecular function than the current generic GO:0019239 and should be added by ISS/ISO (with/from Mmf1p P40185 or human HRSP12 P52758), not an experimental code. Mitochondrial DNA maintenance and related phenotypes should be treated as downstream biological-process consequences, not the core function. The one substantive limitation is the absence of any direct enzymatic assay on Q9UR06 itself, which a heterologous complementation or in vitro deaminase assay would readily resolve.

Artifacts

• OpenScientist mmf2 catalytic arg conservation
• OpenScientist mmf2 evidence matrix
• OpenScientist mmf2 go decision table
• OpenScientist final report
• OpenScientist final report